255 Out Of Bounds & Socket Timeouts

[EnvBroker1]

I had a customer who was getting alerts, but they were "Out of bounds" and "Socket Timeout" errors. Everything was working fine before their systems "crashed hard" as they put it.

My question is, is there a filter in Nagios, so the customer would not get the out of bounds and socket errors, but if it was an actual disk space warning or something to that effect, they would get those.

I suggested they use Outlook and filter out Nagios emails that contained those strings. But I told them, you might want to know when Nagios isn't getting reliable communication from your host machines.

Our Unix guys ended up killing the PID for nrpe and then restating it, so it's communicating again. But the customer (not the Unix shop) was just a little agitated that they kept getting the alerts.

Thanks
Eric

[tmcdonald]

But I told them, you might want to know when Nagios isn't getting reliable communication from your host machines.

I think this is the crux of the problem. They want to hide alerts for when NRPE is not responding, but should still get alerts when it is not responding.

Alerts are only sent/not sent based on the exit code - OK, WARNING, CRITICAL, or UNKNOWN. Nagios doesn't differentiate between different CRITICAL states. That sort of logic could be custom-coded into a new notification command, but this would need to be done on a per-service basis.

[EnvBroker1]

Yes, and that's too much to do, service by service. I'll stick with, if they don't want those specific alerts, they'll have to use the MS Outlook filters to get rid of them.

Thanks

[hsmith]

Are we all right to mark this one resolved and close the thread?

[EnvBroker1]

Yes, you can close this thread

[bwallace]

Glad we were able to help. We'll lock this thread now and feel free to open another should you require assistance with anything else.